Immediate and Follow-up Imaging Findings after Cone-Beam CT–guided Transbronchial Lung Cryobiopsy

Abstract

Purpose

To evaluate findings after transbronchial lung cryobiopsy (TBLC) using intraprocedural cone-beam CT (CBCT) and follow-up chest CT examinations.

Materials and Methods

A single-center, prospective cohort study was performed with 14 participants (mean age, 65 years ± 13 [SD]; eight male participants) undergoing CBCT-guided TBLC between August 2020 and February 2021 who underwent follow-up chest CT imaging. Intraprocedural CBCT and follow-up chest CT images were interpreted for changes compared with baseline CT images. Statistical analyses were performed using independent samples t test and analysis of variance.

Results

A total of 62 biopsies were performed, with 48 in the field of view of CBCT immediately after biopsy. All 48 biopsy sites had evidence of postprocedural hemorrhage, and 17 (35%) had pneumatoceles at the biopsy site. Follow-up CT images showed resolution of these findings. Solid nodules developed at 18 of the 62 (29%) biopsy sites.

Conclusion

Postbiopsy hemorrhage and pneumatoceles on intraprocedural CBCT images (which were clinically occult and resolved spontaneously) and new solid nodules on follow-up chest CT images were commonly observed after TBLC. These findings may help alleviate unnecessary follow-up imaging and tissue sampling.

Keywords: Biopsy/Needle Aspiration, CT, Lungs, Lung Biopsy, Interventional Bronchoscopy

© RSNA, 2023

Summary

Postbiopsy hemorrhage and pneumatoceles were commonly observed at cone-beam CT–guided transbronchial lung cryobiopsy. Follow-up imaging demonstrates these changes resolve spontaneously, though nodules may form at biopsy sites.

Key Points

• ■ Transbronchial lung cryobiopsy frequently caused localized hemorrhage and pneumatoceles as part of the procedure, which were observed at intraprocedural cone-beam CT imaging.

• ■ Following transbronchial lung cryobiopsy, small nodules were often visible on CT images at the site of biopsy and generally decreased in size over time.

Introduction

Transbronchial lung cryobiopsy (TBLC) is an increasingly used minimally invasive technique that yields superior specimen size and quality over traditional percutaneous or transbronchial biopsy techniques (Fig 1) (1,2). TBLC involves the use of imaging-guided bronchoscopy to advance a cryoprobe to the desired biopsy site. Once activated, the metal tip of the cryoprobe is rapidly cooled by flowing compressed gas, creating an ice ball measuring up to 2 cm in diameter. Then, the cryoprobe uses mechanical force to remove tissue through the bronchoscope into a specimen container. This generally preserves the tissue in the specimen without causing substantial damage to tissue structure (3). Traditional transbronchial lung biopsy relies on forceps to sample lung tissue which limits the size of the sample and frequently results in maceration of the specimen, degrading the quality of tissue for pathologic analysis. Thus, TBLC is most commonly used for diseases that require large samples with minimal degradation, such as interstitial lung and diffuse lung diseases.

Figure 1:

(Left) Cryoprobe with a 5-mm ice ball. While ex vivo, the ice ball forms from moisture in the air; in vivo, the ice ball forms within lung parenchyma, supplying the tissue for sampling. (Right) Gross tissue from a transbronchial lung cryobiopsy demonstrates the nidus of tissue with adjacent pulmonary vascular casts.

Intraprocedural imaging with cone-beam CT (CBCT) is a technique used in many minimally invasive procedures whereby a C-arm rotates in a circular motion around a region of interest, capturing two-dimensional images across 200 degrees (4). Software then takes these data and creates a three-dimensional volume across this region without the need for a cumbersome CT machine in the interventional suite. CBCT allows for precise placement of the cryobiopsy probe, potentially decreasing complications such as pneumothorax (5,6). Although others have reported on the technical and clinical aspects of TBLC, little is known about the immediate and late imaging findings of the technique. This study examined the immediate imaging appearance following TBLC with CBCT examinations and the subsequent appearance of the biopsy site with traditional CT scans.

Materials and Methods

Study Participants

This Health Insurance Portability and Accountability Act–compliant study was approved by the Stanford University institutional review board (protocol #59709) with waiver of informed consent. A prospective cohort study was undertaken of 26 consecutive participants undergoing CBCT-guided TBLC at our institution from August 2020 through February 2021. Those without a follow-up chest CT scan in our system were excluded, resulting in 14 participants in the final cohort.

TBLC Procedure

During the procedure, participants were placed supine on a fluoroscopy table and received general anesthesia. A bronchoscope was advanced to the expected location of biopsy with fluoroscopy, and a 1.9-mm cryoprobe (Erbe USA) was placed through the lumen of the bronchoscope. Intraprocedural CBCT was performed to confirm appropriate cryoprobe placement immediately prior to performing TBLC. TBLC was performed using a 6-second freeze followed by en bloc removal of the cryoprobe and bronchoscope with the tissue sample. In the setting of diffuse lung disease, one to four samples were generally taken from different sites in each lobe in a single lung. In the setting of localized disease, two to four samples were taken at the site of abnormality.

CBCT Imaging

Intraprocedural CBCT imaging was performed using the Axiom Artis Zee Fixed Interventional Laboratory (Siemens Healthineers). CBCT images were obtained with a 200-degree rotation lasting 8 seconds using the automated low-dose settings. Images were reconstructed using the standard kernel with 0.75–1-mm section thickness.

Chest CT Imaging

Follow-up chest CT imaging was performed at the discretion of each participant’s primary pulmonologist. The CT images were acquired with or without the administration of contrast material using various scanner models by two vendors (General Electric HealthCare and Siemens Healthineers). Images were reconstructed using lung and soft-tissue kernels with thin sections, 1.25 mm or less. Follow-up CT imaging parameters were in accordance with the institutional standard clinical imaging protocols and varied depending if participants underwent full-, low-, or ultra-low-dose chest CT imaging or pulmonary angiographic CT imaging.

Imaging Analysis

Each CBCT and follow-up CT scan was evaluated by two fellowship-trained chest radiologists (B.P.P. and H.H.G., with 1 and 9 years’ experience, respectively), and any discrepancies were resolved by contemporaneous review and consensus. Clinical and pathologic data were obtained from the electronic medical record, including age, sex, indication for biopsy, biopsy diagnosis from the pathology reports, reported procedural complications, and the final clinical diagnosis from the latest clinical and multidisciplinary conference notes.

Statistical Analysis

Independent samples t test and analysis of variance were performed, as indicated, using Microsoft Excel (Microsoft). A P value < .05 was considered to indicate a significant difference.

Results

Participant Characteristics

Fourteen participants who underwent both CBCT-guided TBLC and a follow-up clinical chest CT scan were included (Table 1). There was a slight male participant predominance, with six female participants (43%) and eight male participants (57%). The average age was 65 years ± 13 (SD) at the time of biopsy. The indication for most biopsies was interstitial lung disease (n = 12, 86%), while two biopsies were performed for potential infection and lymphangitic tumor, respectively. All 14 pathology reports demonstrated at least one histopathologic abnormality, with 12 (86%) showing findings concordant with the diagnosis from the most recent multidisciplinary discussion or most recent clinical note.

Table 1:

Participant Characteristics and Indications

Postbiopsy Imaging Findings

During the 14 procedures, 62 biopsies were performed, with 31 (50%) biopsies occurring in the right lung and 31 (50%) occurring in the left. Of these, 48 of the biopsy sites were imaged immediately with CBCT during the procedure (Table 2). This amounted to a mean of 4.43 biopsies per participant ± 0.93. All 48 biopsies showed ground-glass opacities with or without consolidation at the biopsy site, suggestive of parenchymal hemorrhage (Fig 2). No participant demonstrated any evidence of potentially significant hemorrhage such as hemodynamic instability, hemoptysis, or hypoxia. Pneumatoceles were observed at 17 of the 48 (35%) biopsy sites, measuring an average of 6.9 mm ± 4.3 in the major axis by 4.5 mm ± 3.0 in the minor axis. The number of pneumatoceles was positively correlated with the number of biopsies (analysis of variance r = 0.34, P < .001). One of the 14 participants (7%) developed a small intraprocedural pneumothorax that did not require further treatment. No other major complications were reported.

Table 2:

Imaging Sequelae of Transbronchial Lung Cryobiopsies Performed in Study Participants

Figure 2:

(A) Pretransbronchial lung cryobiopsy CT image shows a small region of ground-glass opacification with traction bronchiolectasis in the left lower lobe (dashed oval). (B) Intraprocedural cone-beam CT (CBCT) image shows accurate placement of the cryoprobe within the same region. (C) Intraprocedural CBCT image shows the cryoprobe within the right middle lobe prior to biopsy. (D) Intraprocedural CBCT image at the same level immediately following biopsy shows regions of ground glass and consolidation, suggestive of hemorrhage (star) and a small postprocedural pneumatocele (arrow).

Follow-up Imaging Findings

Follow-up chest CT imaging occurred an average of 97.8 days ± 63.4 following biopsy (range, 18–200 days). All hemorrhage and pneumatoceles had resolved by the time of follow-up CT imaging. However, solid nodules had developed at 18 of the 62 (29%) biopsy sites (Fig 3), with four occurring at the sites of prior pneumatoceles. The nodules measured on average 6.2 mm ± 3.8 in the major axis by 4.6 mm ± 2.7 in the minor axis and were invariably round or oval with well-circumscribed margins. The presence of nodules was not associated with sites of prior pneumatoceles (P = .26) or length of time between biopsy and CT imaging (P = .43). Multiple follow-up CT scans were available in three participants, demonstrating a progressive decrease in size of all nine nodules initially detected.

Figure 3:

(A) Partially expiratory phase CT pulmonary angiogram in a 41-year-old female participant with a history of breast cancer shows intralobular septal thickening. Concern was raised for lymphangitic spread of tumor versus pulmonary edema. (B) Intraprocedural cone-beam CT shows accurate deployment of the cryoprobe (open diamond) in the region of concern. This biopsy was negative for malignancy. (C) CT image 3 weeks after biopsy demonstrates the development of a solid pulmonary nodule (arrow) at the site of negative biopsy. (D) CT image 5 months after biopsy shows resolution of the nodule.

Discussion

As providers attempt to maximize diagnostic yield while limiting complications, TBLC continues to evolve. Confirmation of TBLC probe placement by CBCT guidance may improve this risk-benefit ratio for potential TBLC candidates. Although the current study was primarily performed for the purposes of understanding immediate and long-term imaging findings from TBLC, our diagnostic yield in terms of reporting histopathologic abnormalities (100%) and concordance with multidisciplinary discussion diagnosis (86%) is consistent with the literature on TBLC (7–9). Additionally, the 7.1% pneumothorax rate in this sample is lower than the 25% reported in a larger study (9).

It is important to note that intraprocedural CBCT images revealed hemorrhage after biopsy that was not apparent at bronchoscopy. To our knowledge, no other study has reported the incidence of radiographically visible hemorrhage directly after cryobiopsy with CBCT. Zhou et al (9) noted a 75% rate of bronchoscopically visible local hemorrhage, with the majority being considered mild. Therefore, bronchoscopic evaluation for postbiopsy bleeding may underestimate the degree of intraparenchymal bleeding present. Although bleeding is subclinical, focal parenchymal hemorrhage could obscure focal areas of disease involvement, complicating probe placement or pathologic assessment at repeat biopsy and potentially worsening gas exchange abnormalities in patients with decreased physiologic reserve.

To our knowledge, pneumatocele creation at the site of TBLC has also not been previously reported. Although pneumatoceles were only detected in 35% of biopsies in the current study, it is conceivable that other pneumatoceles in the remaining biopsies were masked by local postbiopsy hemorrhage. Intraprocedural pneumatoceles were not visible at follow-up imaging and were not associated with postbiopsy nodules, suggesting that they spontaneously resolve. The development of nodules at 29% of biopsy sites is comparable with the reported 30% rate in traditional transbronchial biopsies (10). These nodules presumably represent small clots and/or granulation tissue. Characterization of these nodules as expected postbiopsy changes is important to avoid misinterpretation at follow-up imaging given the increased risk of lung cancer in patients with interstitial lung abnormalities (11,12). In the current study, nodules were present at the biopsy site for as long as 200 days after biopsy, but importantly, all nodules demonstrated regression in size and number at multiple subsequent imaging studies.

Limitations of this study include not having all biopsy sites imaged with CBCT after biopsy, limiting assessment for immediate postbiopsy complications at these sites. The relatively small sample size may somewhat limit the generalizability of this study. While in theory, nodules could have developed separately from the biopsies, we felt the likelihood was low as the locations of these new nodules could be definitively traced back to the biopsy site on the intraprocedural images. Finally, the lack of a set follow-up time for the postprocedural CT imaging was suboptimal for assessing the presence and size of nodules, as participants with multiple follow-up chest CT scans showed a progressive decrease in nodule size over time. However, the length of time between biopsy and follow-up CT imaging were not significantly associated in this study.

In summary, postbiopsy hemorrhage and pneumatoceles are frequently found immediately after TBLC but resolve spontaneously. However, nodules frequently form at the site of biopsy and should not be mistaken for a new neoplastic process. Intraprocedural CBCT can improve reader confidence that these nodules are at the site of prior biopsies.